On September 29, 2026, Reuters published an investigation into how McDonald’s prices your Big Mac. According to the report, McDonald’s runs a machine-learning pricing engine across its nearly 14,000 U.S. restaurants. The engine analyzes data from millions of daily transactions to generate what the company calls the optimal price at each location. One of the inputs is an estimate of how much each store’s customers are willing to pay. Reuters found that a company-run store in Fresno, California sold a Big Mac for $5.69. Another company-run restaurant two miles away sold the same sandwich for $6.89, a 21 percent premium. Reuters could not confirm the engine caused that specific gap.
McDonald’s pushed back firmly. The company said the pricing portal is a tool, not a mandate, and that franchisees always determine the final price. It called the reporting speculative and uninformed. Both of those things can be true at once. The AI can be a recommendation engine rather than a mandate, and the outcome can still be two prices for the same Big Mac two miles apart. That is exactly what makes this case instructive for every retailer deploying AI-driven pricing.
The AI did what it was designed to do. It optimized price by location based on willingness to pay. The technology worked exactly as intended. What the McDonald’s case exposes is not a technology failure. It is a governance gap. When an AI pricing engine optimizes for margin without a governance framework that defines the limits of acceptable price variation, it produces outcomes that are mathematically optimal and reputationally damaging at the same time. The lesson is not that AI pricing is dangerous. It is that AI pricing without governance is.
What the AI Actually Optimized For
The screenshots Reuters reviewed reveal the mechanism precisely. The franchisee interface displays messages like the restaurant is showing MEDIUM SENSITIVITY to price, based in part on customer willingness to pay in the area. The engine, operated by Tiger Analytics with McDonald’s supplying the rules and corporate targets, incorporates public pricing data from nearby competitors like Wendy’s and Burger King. It then generates a recommended price calibrated to what the local customer base will tolerate.
This is textbook price optimization. The AI is doing sophisticated, legitimate work. It estimates demand elasticity at the store level and recommends the price that maximizes revenue. In a spreadsheet, this is exactly what a pricing team would want. However, willingness to pay is not evenly distributed, and neither is it neutral. A store in a lower-income neighborhood with less competition may show lower price sensitivity. That is not because those customers value the product more. It is because they have fewer alternatives. The AI reads that as an opportunity to charge more. The optimization is correct. The outcome is a higher price for the customers with the fewest options.
As I described in the retail AI decision gap analysis, the retailers deploying AI inside the decision itself gain real speed and margin advantages. McDonald’s pricing engine is exactly that: AI inside the decision. But this case shows the other edge of that same capability. When the AI makes the decision, the AI also makes the mistakes, at machine speed and machine scale, across 14,000 locations, unless a governance layer constrains what it optimizes toward.
The Three Governance Gaps the Case Reveals
No Fairness Constraint on the Optimization
The first gap is the absence of a fairness constraint. An AI pricing engine optimizing purely for revenue will charge the maximum each location will bear. A governance framework would define guardrails: a maximum acceptable price variation for the same product between nearby stores, a floor and ceiling relative to a regional average, or an explicit rule that willingness-to-pay signals correlated with limited consumer choice cannot raise prices. None of these constraints reduce the AI’s capability. Instead, they define the boundaries within which that capability operates. Without them, the engine optimizes into exactly the outcome that generates a Reuters investigation.
No Clear Ownership of the Final Decision
McDonald’s says franchisees set their own prices. Yet five owners told Reuters they felt pressured to follow the AI recommendations. Since January, franchisees have had to engage constructively with the approved pricing tools. The CEO reportedly told investors that pricing non-compliance factors into franchisee business reviews. That is a governance contradiction. The framework says the human owns the decision, but the incentive structure pushes the human to defer to the AI. When the stated ownership and the actual incentive point in different directions, the AI effectively makes the decision while the human carries the accountability. As I described in the Anthropic commerce agent blueprint analysis, the retailers who get AI governance right define clear decision ownership and then align the incentives to match it. A recommendation the human is pressured not to override is not a recommendation. It is a mandate with deniability.
No Reputational Circuit Breaker
The Connecticut franchisee whose pricing tool suggested roughly $18 for a Big Mac meal is the clearest example of the third gap. A governance framework would include a circuit breaker. That means a threshold beyond which a recommended price gets flagged for human review before it ever reaches the menu. A price that extreme is a reputational risk regardless of whether the local demand data supports it. The $18 Big Mac meal went viral. The math may have been defensible. The brand damage was not. A circuit breaker exists to catch the outputs that are locally optimal and globally damaging. The McDonald’s case shows what happens without one.
Why This Matters for Every Retailer, Not Just McDonald’s
Dynamic AI pricing is spreading across retail. Grocery, fuel, travel, and general merchandise are all moving toward location-level and even individual-level price optimization. The capability that McDonald’s deployed is available to any retailer with transaction data and a pricing engine. Consequently, the governance gap the McDonald’s case exposes is not a McDonald’s problem. It is a preview of the decision every retailer deploying AI pricing will face.
Furthermore, the regulatory environment is shifting underneath this capability. The Reuters report noted that the pricing portal’s own terms of service warn that franchisees may be competitors under antitrust law. When an AI engine coordinates pricing across nominally independent operators, the antitrust exposure is real and growing. As I described in the AI shopping fraud and trust analysis, regulators are beginning to scrutinize AI-driven commerce decisions. The retailer who has a documented governance framework in place is in a far stronger position than the one who deployed the capability and hoped the outcomes would be defensible.
What This Means for LatAm Retailers
Dynamic pricing is arriving in LatAm retail, and the governance considerations are more acute for two reasons. First, income inequality is more pronounced across many LatAm markets. A willingness-to-pay optimization therefore has a higher risk of producing prices that track a neighborhood’s economic vulnerability rather than genuine product value. An AI engine that charges more where consumers have fewer alternatives produces sharper disparities in markets with wider income gaps. Second, consumer trust in retail pricing fairness is already fragile in several LatAm markets with histories of price volatility. A dynamic pricing deployment that produces visible disparities risks damaging a trust relationship that is harder to rebuild than in more stable markets.
However, this raises the value of getting the governance right from the start. The LatAm retailer who deploys dynamic pricing with explicit fairness constraints, clear human ownership, and reputational circuit breakers gains the margin benefit while protecting the trust relationship. That relationship is the foundation of long-term customer loyalty. The governance framework is not a constraint on the value of AI pricing. It is what makes that value sustainable.
The Question Every Retailer Deploying AI Pricing Should Answer
What Is Your AI Allowed to Optimize Toward?
The McDonald’s case comes down to a single unanswered question. What was the pricing engine allowed to optimize toward, and what limits applied to that optimization? Suppose the answer is that the engine was told to maximize revenue per location with no fairness constraint, no true human ownership, and no reputational circuit breaker. Then the 21 percent gap and the $18 Big Mac meal were not malfunctions. They were the system working as designed. Nobody asked the AI to consider whether the optimal price was also a defensible one.
Governance Is the Product Decision, Not the Compliance Afterthought
The retailers who deploy AI pricing successfully over the next few years will treat governance as a core part of the deployment. They design it before the engine goes live, rather than as a compliance response after a headline. The fairness constraints, the ownership structure, and the circuit breakers are not features that slow the AI down. Instead, they keep the AI’s output aligned with the retailer’s long-term interest rather than just its next-quarter margin. Optimization without governance optimizes for the wrong time horizon.
The AI Optimizes Whatever You Point It At
McDonald’s built a pricing engine that does exactly what sophisticated price optimization should do. It found the optimal price at every location. But optimal for margin and optimal for the business are not the same thing once you account for consumer trust, franchisee relationships, and regulatory exposure. The technology worked. Governance did not keep pace. Every retailer deploying AI pricing is about to face the same choice McDonald’s faced. Define what the AI is allowed to optimize toward, and what it is not, before the engine makes those decisions for you at the scale of your entire network. The AI will optimize whatever you point it at. Governance is how you make sure you pointed it at the right thing.
If you are deploying AI-driven pricing or building the governance framework that keeps price optimization aligned with consumer trust and regulatory compliance, connect with me here or reach me on LinkedIn. I am happy to walk through the framework we use across the U.S. and Latin America.
Adriana Rivas is a retail technology executive and AI strategist. She is the recipient of the Gold Stevie® Award, Thought Leader of the Year 2026, recognized by Thinkers360 as the #7 Global Thought Leader in Retail, and named to the RTIH Top 100 Retail Technology Influencers 2026. She is the author of How to Implement Self-Service Without Failing, now available in a Revised and Expanded Edition.